Jump to content

UE 5.1 nanite foliage (No LODS)


HoBGoBLiNzx3
 Share

Recommended Posts

This isn't a proposal to change graphics engines, so chill. Imagine DCS implementing a version of this for it's own engine.. If they ever drop a new game title from scratch, they will have so many awesome features to draw from. No more LODS?? No more constant shadows changes.. This is crazy. 

However if they do release a DCS World 2.0 and decide to go with UE, and port over all their physics calculations somehow.. It will be insane.

As stated in the video the recommend environment in 5.0 and prior was 22KM.. Now, it's 88,000,000km. This is absolute insane. 

You can put this where it needs to go admin, I didn't really know where to post it. 

VR

 


Edited by HoBGoBLiNzx3
  • Like 3
Link to comment
Share on other sites

2 hours ago, HoBGoBLiNzx3 said:

However if they do release a DCS World 2.0

More like 4.0 (which I'm sure will happen eventually if the company remains solvent). We're on 2(.8) now, Vulkan/MT will be 3+. 

  • Like 1
  • Thanks 1

Слава Україні, слава героям!

Link to comment
Share on other sites

On 11/20/2022 at 9:09 PM, Mars Exulte said:

More like 4.0 (which I'm sure will happen eventually if the company remains solvent). We're on 2(.8) now, Vulkan/MT will be 3+. 

It's still DCS WORLD version 2.8

I meant DCS WORLD 2  vs 1.0 etc.

Link to comment
Share on other sites

Should please NOR

System: 9700, 64GB DDR4, 2070S, NVME2, Rift S, Jetseat, Thrustmaster F18 grip, VPC T50 stick base and throttle, CH Throttle, MFG crosswinds, custom button box, Logitech G502 and Marble mouse.

Server: i5 2500@3.9Ghz, 1080, 24GB DDR3, SSD.

Link to comment
Share on other sites

9 hours ago, HoBGoBLiNzx3 said:

It's still DCS WORLD version 2.8

I meant DCS WORLD 2  vs 1.0 etc.

There is no point in making "DCS World 2". We're on DCS 2.8 now, will be on 3.0. Unreal is nice, but it's not a bespoke engine made for flight simulation. I wish people stopped suggesting an engine change as a magical cure-all for all issues with DCS. An engine change is coming, not to a commercial offering, but to a custom-made Vulkan renderer. This should solve the current performance issues, improve VR, and generally make things better. 

UE5.1 is not the magic bullet. If someone wants to make a flight sim on that engine, they're free to try. However, there's no point trying to port DCS to it. I know pretty videos are pretty, but there's more to DCS than that.

Although, perhaps they could borrow some tech.

Link to comment
Share on other sites

5 hours ago, Mr_sukebe said:

Should please NOR

Yep.
Such a shame it's not a commercial product... 😕

 

 PERFORMANCE ISSUES WITH DCS ?  ==> DCS 2.5.6 <==                                          aka "Ducko"  -  my mods/files

LM transparent DCS icons_hs_LESSshdw - Sep29_2022_smaller.png
Win10 Pro x64 | Intel i7 10700KF (all core OC@4.9) | 64GB DDR4 - Crucial Ballistix (OC@3466/CL16) | RTX 3090 24GB - EVGA FTW3 Ultra (UV@1800/1250/812mV) | 1TB NVMe + 500GB SSD + 4TB TB HDD | Corsair RMX 850W | 
MSI Z490-A Pro | Fractal Meshify C tg | 43'' 4K QLED Toshiba 43QA4C63DG | M-Audio USB + Sennheiser HD-599SE | TP-Link Hub 7x USB 3.0 |  VKB NXT stick + Logitech X56 throttle | TrackHat Clip w/ OpenTrack (Head Tracking)

Link to comment
Share on other sites

17 hours ago, HoBGoBLiNzx3 said:

It's still DCS WORLD version 2.8

I meant DCS WORLD 2  vs 1.0 etc.

Yeah, I see what you mean. But they're not doing a whole new release with new product branding (which is all that means). When I first discovered DCS it was 1.2, they went to 1.5 with a new engine. 2.0 (up through our current 2.8) was a whole new engine. 3.0+ will be Vulkan with multithreading.

Basically each of these major releases IS exactly what you're describing, a whole new engine, but your previous purchases are still good and preexisting assets are ported over (which is also what you mentioned wanting). By strict ''engine releases'' technically we are on the third major engine/api release now with the fourth inbound since the ''DCS World'' brand was launched. They just don't append a number to the title, that's all, it's in the version number.


Edited by Mars Exulte

Слава Україні, слава героям!

Link to comment
Share on other sites

Whatever the engine, it would be nice to have similar tech available on our side to do away with LODs in DCS at some point in the future. Tree and shadow pop-in we've got now (especially on Marianas map) really looks meh.

Maybe it would also help with neverending target spotting issues, 'cause these are currently LOD related to some margin as well.


Edited by Art-J
  • Like 1

i7 9700K @ stock speed, single GTX1070, 32 gigs of RAM, TH Warthog, MFG Crosswind, Win10.

Link to comment
Share on other sites

On 11/22/2022 at 5:27 PM, Mars Exulte said:

Yeah, I see what you mean. But they're not doing a whole new release with new product branding (which is all that means). When I first discovered DCS it was 1.2, they went to 1.5 with a new engine. 2.0 (up through our current 2.8) was a whole new engine. 3.0+ will be Vulkan with multithreading.

Basically each of these major releases IS exactly what you're describing, a whole new engine, but your previous purchases are still good and preexisting assets are ported over (which is also what you mentioned wanting). By strict ''engine releases'' technically we are on the third major engine/api release now with the fourth inbound since the ''DCS World'' brand was launched. They just don't append a number to the title, that's all, it's in the version number.

 

Yeah no doubt. That's why I said if they ever do.. At some point, older hardware wont be able to run the newer games and the content you bought has to be able to be played somehow. I'm guessing use it as a free legacy build maybe? I seriously wouldn't be opposed to purchasing the actual game with 4-6 maps for like 120 bucks if it had something like UE5.2 and onward. Mostly because I know that whatever hardware I buy, will have a real world effect on the quality and performance. Small limitations aside. I'm not a programmer but at some point these technologies will have to all be made available because the way technology moves, we could end up with something that is 15 years old while every other new game is getting brand new modules that plug right into the engine.. Which seems to kind of be the reason why a lot of companies use these main engines. They have so much support... It would free up so much time I'm guessing for the developers to come up with content.. Also, a sandbox editor so that custom maps could be made, downloaded, and shared. Though I think this is more of a business decision more than a logistical one. While the switch would take a couple years, everything post release would be much faster, easier, and sellable content would come out much quicker. But hey, that's just my perspective from someone who doesn't work in the industry.. 

Link to comment
Share on other sites

On 11/22/2022 at 9:33 AM, Dragon1-1 said:

There is no point in making "DCS World 2". We're on DCS 2.8 now, will be on 3.0. Unreal is nice, but it's not a bespoke engine made for flight simulation. I wish people stopped suggesting an engine change as a magical cure-all for all issues with DCS. An engine change is coming, not to a commercial offering, but to a custom-made Vulkan renderer. This should solve the current performance issues, improve VR, and generally make things better. 

UE5.1 is not the magic bullet. If someone wants to make a flight sim on that engine, they're free to try. However, there's no point trying to port DCS to it. I know pretty videos are pretty, but there's more to DCS than that.

Although, perhaps they could borrow some tech.

It's a graphics engine, why would it NOT work? If they can build 40,000,000 km detailed worlds, with realtime raytracing, shadows, lumen, and all the features that come with it at refresh rates that we currently have... why would it NOT? Unreal has 200x more support and is constantly updated..  If an aging DCS engine that doesn't even have multicore support or vulkan can run it.. there is absolutely no reason why UE5 can't..  

That wasn't really my intention in the first place. It's the technology that is available and the quality they are achieving.. It would be nice to have.. but instead we wait for mc and vulkan lol.

DCS isn't the only simulator in the world lol, in fact it's one of the smaller simulators on the market... in fact it almost never makes top 20 sim lists (not because it's not good by any means) but to assume that DCS is somehow special and engines that AAA studios use just couldn't handle a small market, small developer, simulator doesn't seem right to me... 

However, Like I said. I don't work in the field, so if you do. Please share why this would be, I'm all ears and always willing to learn... However if it's just some opinion you read on the forums then, I'll continue with my doubts. 

Link to comment
Share on other sites

4 hours ago, HoBGoBLiNzx3 said:

It's a graphics engine, why would it NOT work?

Because a game is not just a graphics engine. It's way more than that, UE is an entire framework on which a game is built. DCS is too tightly interwoven with its engine that it would not be possible to port without starting from scratch. Besides, we don't know what kind of tricks UE does in order to achieve those numbers. The videos you posted are advertising, but how well does this fly in a real environment? It tells you all about things it can do, but nothing about things that would have to be worked around, because the engine assumed that things will be happened in a given way. For example, in Satisfactory (a game built on an earlier version of UE), there are no conveyor belts that move more than 780 items per second, because the physics engine would break if an object moved any faster than on the current fastest belts. Bad news for a sim that needs its aircraft to be able to exceed Mach 2 and missiles to hit things while moving at more than Mach 4. UE's physics engine is not made for flight simulation or for high supersonic flight, in contrast to DCS, which is made for flying from ground up, and while it might have other limitations and tradeoffs, it has them in places where they don't matter, just like an UE5.1 game may correctly assume nobody is going to run around at Mach 3.

I don't work in the field, but I have a brother who went into game dev, and I was once a semi-prominent mod developer on two separate engines. So I've had a chance to get acquainted with technology and its limitations, and with how much work it is to deal with them if you want to push those limitations (let's just say Unity isn't the best engine to build realistic orbital mechanics in). The "top 20" sim lists aren't exactly useful, BTW, because they either have a very loose definition on a sim (such as lumping WT and Ace Combat in), or they include other genres such as car sims, which are much easier to figure out than flying. If you wanted to make a "top 20 flight sims" list, then you'd run out of modern ones halfway through and have to fill the rest with oldies from the era of Jane's F-15. There are exactly four non-pro combat sims in active development right now, of which two are WWII-only and far simpler than DCS is.

Link to comment
Share on other sites

Dragon>
saying that it would be impossible to port DCS to work with U5 without a complete rebuild is a very strong statement for someone is neither a developer or works for ED.

We know ED have been working on this for a long time.  We’ve also seen examples like NOR which is using Unreal for a flight sim engine.

For all we know, it’s possible that ED have been beavering away on the game logic, such that it can plug into the Unreal engine when U5 supports VR (as far as I’m aware, that’s currently not the case).

 

The reality is that we simply don’t know, and actually, I’d rather ED don’t tell us the plan, as we’ve seen far too many times the amount of “noise” that is created when ED pencils in a potential date and misses it.  It’s a LOT more sensible for them to wait and the surprise us when it’s ready.

System: 9700, 64GB DDR4, 2070S, NVME2, Rift S, Jetseat, Thrustmaster F18 grip, VPC T50 stick base and throttle, CH Throttle, MFG crosswinds, custom button box, Logitech G502 and Marble mouse.

Server: i5 2500@3.9Ghz, 1080, 24GB DDR3, SSD.

Link to comment
Share on other sites

2 hours ago, Mr_sukebe said:

saying that it would be impossible to port DCS to work with U5 without a complete rebuild is a very strong statement for someone is neither a developer or works for ED.

It's a completely true one, which will be obvious for any programmer that had worked with game engines. Fundamentally, it's a matter of certain programming paradigms that DCS follows and UE does not (it follows different ones). Even upgrading to the next version of the same engine is often painful. DCS devs are working on their own solution, and whatever it will be, we'll get it in due time. UE will not be the panaceum to allay all our performance and eyecandy woes.

2 hours ago, Mr_sukebe said:

We’ve also seen examples like NOR which is using Unreal for a flight sim engine.

We haven't seen "examples", we have heard of an example, in development right now, that seems to be aimed at a completely different market. The problems are, in principle, solvable, one can substitute a custom physics engine, for instance, which would not suffer from the speed limitations Satisfactory suffers from. The question is, are they worth solving? NOR team seems to have decided that yes, they are, and they have resources to try. Whether they succeed, whether their sim is any good, and whether it is even available to anyone but governments, it all remains to be seen. Also note that professional sims such as NOR can, for example, completely eschew AI in favor of being multiplayer-only, in effect. In a training environment, you can get together a large number of people on a single LAN, and you can afford to have every single aircraft piloted by a human, with the exception of transports and the like which can run on a simple autopilot. OTOH, they will usually have advanced networking features enabling interaction with other virtual training environments such as VBS. 


Edited by Dragon1-1
Link to comment
Share on other sites

Apologies, but your utter intransigence about it being "impossible" basically prevents any further discussion on the subject.

 

As for only having "heard" of an example, I bring forward the following witness for the defence:

Don't bother replying.

System: 9700, 64GB DDR4, 2070S, NVME2, Rift S, Jetseat, Thrustmaster F18 grip, VPC T50 stick base and throttle, CH Throttle, MFG crosswinds, custom button box, Logitech G502 and Marble mouse.

Server: i5 2500@3.9Ghz, 1080, 24GB DDR3, SSD.

Link to comment
Share on other sites

7 minutes ago, Mr_sukebe said:

Apologies, but your utter intransigence about it being "impossible" basically prevents any further discussion on the subject.

No, it doesn't. See? I can still discuss it, and there's not a damn thing you can do about it. Also, "impossible without a major rework". Of course making a new sim is possible, and of course, if they were willing to expand an amount of work comparable with making a new sim, they could "port" DCS to UE. It would save no effort, in fact, it would require more effort than adding multicore and Vulkan to DCS engine. 

As for NOR, I see a two minute tech demo with many pretty pictures and no action going on. This is something that's easy enough to throw together an an early stage of a project. What we don't see is actual flying and fighting (dogfight mode in the Viper looks OK, at least, so at least avionics are partially done). Let me know when they have a video of a full mission flown, if not from startup to shutdown, then at least from ingress to egress. I don't doubt they'll get there, but I can bet that we'll get multicore in DCS before we get that.

Link to comment
Share on other sites

Well, looking at the racing sims side of the industry, Kunos Simulazioni somehowe managed to combine UE4 engine mostly for gfx side with some of their own code for physics side in ACC. Never investigated closer how they managed to do it, 'cause I don't own that actual sim and don't plan to, having enough fun with its predecessor.

It's worth noting, however, that despite good results of this combination and commercial success of the ACC, for their next project of AC2 apparently they're dumping UE and going back to development of their own next gen proprietary engine again, because they said UE one was not flexible enough, whatever they mean by that.

It just shows that no matter what flashy adverts Epic publishes about UE5, it might not be "be-all and end-all" engine well adaptable for all game types.

i7 9700K @ stock speed, single GTX1070, 32 gigs of RAM, TH Warthog, MFG Crosswind, Win10.

Link to comment
Share on other sites

That's exactly what I'm talking about. UE5 might be pretty adaptable, but in the end, if you're going to write your own physics, your own AI, etc., then why should you bother paying for UE5 license? For eyecandy? Can also be done in a bespoke engine. There's absolutely zero point in using UE if you're then going to discard most of it in order to make your own stuff work. I don't know how NOR is doing it, maybe it's found a way to deal with it, or maybe they haven't learnt their lesson yet and will moan and complain at one point how going with UE was a lousy idea, but they've invested so much in that paradigm that they can't reasonably roll their own. I wish UE fanboys would understand that, a lot of people get hooked on meaningless advertising and then demand serious devs spend time and money on an engine wholly unsuitable for the job.

Interestingly, ACC says it supports VR, but judging by Steam reviews, it's got serious performance problems with it. In other words, just like DCS. Those who thinks UE is the solution to their VR performance woes should take a rain check on that. UE5.1 might be better for VR, but so will DCS 3.0, when it comes. 

Link to comment
Share on other sites

Yup, LODs themselves aren't bad, but they do require some more finesse that the current DCS implementation features. Although, the biggest LOD problems are with models that are themselves severely outdated.

From their videos, NOR doesn't look any better or worse than DCS on high graphics settings. It's UE5, so it'll support multithreading and Vulkan from the start, but I wouldn't be surprised if performance was similar once that makes it in.

Link to comment
Share on other sites

 

Somehow, they all miss an important feedback function 🕹️

 

 

Gigabyte Aorus X570S Master - Ryzen 5900X - Gskill 64GB 3200/CL14@3600/CL14 - Asus 1080ti EK-waterblock - 3x Samsung 980Pro 1TB - 1x Samsung 870 Evo 1TB - 1x SanDisc 120GB SSD - Heatkiller IV - MoRa3-360LT@9x120mm Noctua F12 - Corsair AXi-1200 - TiR5-Pro - Warthog Hotas - Saitek Combat Pedals - Asus PG278Q 27" QHD Gsync 144Hz - Win11 Pro/Linux - Phanteks Evolv-X 

Link to comment
Share on other sites

vor 19 Stunden schrieb HoBGoBLiNzx3:

Yeah then you hit 2:00 on when he explains what I meant... 

It's what they say it will be able to do, once finished...and likely dozens and dozens of patches later.

 

Advertising videos are worth NULL.

I grew up with TV-Ads of the `70s...cmon...at the age of 6 you realize there can't be a more white washing detergent every week and nothing is as advertised, likely the other way around tbh.

Second, I don't think you can launch such a simulation from zero-to-full_fidelity on Day-1.

 

Proof me wrong and we are all happy, maybe not all but you get my point.

Gigabyte Aorus X570S Master - Ryzen 5900X - Gskill 64GB 3200/CL14@3600/CL14 - Asus 1080ti EK-waterblock - 3x Samsung 980Pro 1TB - 1x Samsung 870 Evo 1TB - 1x SanDisc 120GB SSD - Heatkiller IV - MoRa3-360LT@9x120mm Noctua F12 - Corsair AXi-1200 - TiR5-Pro - Warthog Hotas - Saitek Combat Pedals - Asus PG278Q 27" QHD Gsync 144Hz - Win11 Pro/Linux - Phanteks Evolv-X 

Link to comment
Share on other sites

DCS World will never be ported to UE5 (or any other version of any other engine) It's not going to happen. 

However, at some point someone will definitely make a completely NEW commercial combat flight simulator that will for all intents and purposes be better than anything else on the market. It's inevitable. With VR on the uptake, it is going to happen. It might even be Eagle Dynamics who do so. It might be in 5 years or 10 but it will happen. 


Edited by Lurker

Specs: Win10, i9-9900KF@5Ghz, 32GB DDR4 RAM 3200XMP, 1 TB M2 NVMe SSD, KFA2 RTX3090, VR G2 Headset, Warthog Throttle+Saitek Pedals+MSFFB2  Joystick. 

Link to comment
Share on other sites

 Share

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...