Cyberpunk 2077 Is Amazing... So Why Switch To Unreal Engine 5?
Vložit
- čas přidán 7. 10. 2023
- ► Watch the FULL video here: • DF Direct Weekly #131:...
► Support us on Patreon! bit.ly/3jEGjvx
► Digital Foundry CZcams: / digitalfoundry
► Digital Foundry at Eurogamer: eurogamer.net/digitalfoundry
► Follow on Twitter: / digitalfoundry
If CDPR is never again using their engine and abandoning its development, I would love them to open source it if possible.
Yeah I've been throwing that out there on the CDPR forums for the last year ...... If nothing else Modders could come up with decent expansions to Witcher 3 and Cyberpunk 2077 to hold everyone over for the next 3-4 years before they come out with a new Witcher for Cyberpunk .... For better or worse a lot of the hype surrounding Cyberpunk came from those still modding and player Witcher 3
If nothing else let some of the more experienced modders fix all the issues there still is with Witcher 3 Next Gen which really could and should run better than it does right now
i'm surprised they haven't spun red engine off as a competitor to unreal engine that's managed by a separate team and has an extremely low commision rate of 1 to 5 %.
@@QH96 probably because it is quite a handful to manage, hence their decision to move to UE5 to free up resources (cut costs), there aren't any real arguments in this video apart from some very general testing which means very little, RED engine might be at an end of its development path, they simply cannot make any improvements and are stuck with probably quite expensive dev team to maintain it (PT and RT are slap ons, hence why they perform so poorly on not top of the line hardware, which vast majority of players do not actually have), freeing those assets up would allow for more actual content to be made, instead of continuously putting one bandaid over the previous one on RED
CP2077, despite the looks, had very little actual story content, the main story can be finished in couple hours, it is very short, and the side stuff seems very repetitive, this is a big downgrade from W3
I hope they realize that super complex and difficult to navigate multilayer city is a bad design, could have easily increased the area at the cost of some useless complexity there, but then the game would feel even shorter when you can get to places by being familiar with surroundings (like in W games) instead of looking at the "gps" trace on minimap to try and find a way to another "layer"
They probably can’t open source it, in house engines usually have a lot of licensed tech in them.
@@dsfs17987 While the main story is short, the open world is a upgrade from The Witcher, although they are totally different in concept so can't really be compared, but in no way is the world of cyberpuk a downgrade from that of the Witcher 3.... that game has the same issue red dead has..... 70% of your game is spent going from place to place in quite a boring way.... on a horse.... cyberpunk is deus ex on a huge scale and is quite the achievement in my eyes....
Cyberpunk has little to no stutter, man I'm going to miss this engine.
The only stutter I get in Cyberpunk 2077 is when the game does an autosave while I’m driving a vehicle.
But I can't say the same for the Starfield, it's a mess on my PC and just shooting a normal submachine gun with normal ammo against a empty wall outside combat or even to the sky my FPS can go from 70 to 20 FPS and the effects are just average and there's also no realistic bullet physics at all, something is quite wrong with the game...
@@guily6669The bullet physics thing is a nitpick, but I do agree on the FPS issues. That's not stuttering though
@@guily6669 That has nothing to do with this. Starfield isn't on Unreal Engine 5 nor Redengine or whatever it's called.
PS. You must be doing something veeery wrong if starfield has that massive fps drops for you. The only time the fps goes below 60 on my rx 6600 is in the cities, and that's because of the CPU.
You can disable that with a mod@@Cats_Are_Scary
I read that CDPR are working with Epic Games to bring features they have with Red engine to Unreal engine. So hopefully it will be great all round.
Unreal Engine can be forked and modified to perform specific functions not provided through basic scripts or blueprints. I doubt the Unreal Engine version of Cyberpunk we'll be playing will be running the base version of Unreal
A logical consequence
@@EliteInExileexactly. This is huge for Epic because it's basically taking all that experience and tech from the RED Engine to improve the shortcomings of Unreal Engine.
Unreal 5.4 already has path tracing
5.4 isn't even out yet
Something people rarely talk about, is the fact that an in-house engine poses a lot of risks and workflow barriers when senior devs leave / when the dev team is expanded and new hires are brought in. These new hires will have to be brought up to speed from scratch because nobody can be expected to have any prior experience with an inhouse engine like RED engine - and if you're in the middle of a large scope project, you don't want to suddenly have half your staff consisting of new hires with zero knowledge of the capabilities and limitations of the engine they're using.
Here, the move to Unreal Engine makes a lot of sense. It ensures that should it be needed, CDPR can rely on new devs to have UE5 experience already regardless of their background prior to being hired into CDPR. It opens up the gates for much more fluent project pipelines with fewer risks and hickups. Also noteworthy that CDPR has opened the new studios in the US, meaning lots of new hire is expected/already in place - so the move to UE5 makes a lot of sense here. These new divisions will not have to teach all the new devs everything from the ground up, they can instead be put straight to work at a reasonably high efficiency level - which was something that has plagued CDPR before, as well as any other company with heavy reliance on in-house engine (like EA and Dice who botched Battlefield 2042 because most of the senior devs had left and all the new hires had no idea how to properly make use of the Frostbite engine).
Good answer.
Spot on. Companies with proprietary tools can often times use them as a marketing gimmick. No way of knowing whether their tech is actually good, because it's not accessible to the public. Unreal Engine is available to everyone, so the engineers can easily see where it's good and where it's terrible.
This is why DICE had an entire development team just working on the Frostbite engine to turn it from a small in-house wonder with its own quirks and no documentations, into a cross-studio major engine for racing games, sports games, RPGs, FPS games, etc.
This is also, in part, why Dragon Age Inquisition had a troubled development. They switched to Frostbite very early on and basically had to build almost everything from the ground up with little to no documentation.
@@sujimayneevery Bioware game with Frostbite has a lot technical issue. I think this engine had done more harm to EA in general.
@@SuperhotdogZz I was about to comment on this too. Mass Effect Andromeda devs had a REAL trouble with FrostBite and that certainly hurt the final game a lot in the end.
This is exactly what I'm wondering. Playing this game at 2k using 3080ti, DLSS on Quality, RTX and everything else on highest settings... Except for path tracing and cascade shadows. This game might be the best looking game I've ever played. It certainly has a lot to do with the art direction but God Damn!!!!! This game looks amazing and runs beautifully! On a 65' OLED, it's jaw dropping. More than any other game, I just stop and gasp at this damn game. My wife is sick of hearing me say, "baby... look at this".
I'm running it on a 1660ti at 1080P with most things on high and get a fairly consistent 50-60fps. It's unbelievable how well it runs and looks on this basic of a machine. It is BY FAR the framerate to visuals ratio champion. I get like 25fps in Starfield and it looks like trash for example.
@@mezzb honestly doing similar settings (basically Hardware Unboxed's 2.0 recommended settings) on a 3070 & 4K at DLSS Balanced or 2K at DLSS Quality it absolutely flies. It's even perfectly playable at DLSS Quality at 4K & it looks fantastic - just not quite butter-smooth.
Only issues are the VRAM limit making a few blurry textures hang around now & then if I suddenly drive into a busy crowded scene, but otherwise the game doesn't miss a beat. So glad I waited for 2.0 before diving in.
Funny how the more realistic the graphics get the more you notice how bad the character animation really is.
@@10secondsrule which is pretty nuts considering it's motion capture, the closer to lifelike it gets, the more our lizard brain dislikes what's still obviously fake. Which is why more stylised visual designs, like 3D cartoons, look better.
This has been my biggest concert as a fan of the red engine ever since they announced that they were switching. I have no technical background, so it was just amazing for me to see that ever since the witcher 2 came out, my mid-build pc could run the game smoothly, and the same pc in 2015 could run the legendary witcher 3 game. That was mind-blowing for me, as other games of the same generation struggled to run at the same speed on my PC. now given the introduction of Unreal 5, I can understand why it may work on maybe a witchers game the transition, but for future cyberpunk games, let us just hope for the best.
I think the reason for the switch is tool set and easy of hiring people already trained on the engine, what allows them to kickstart multiple projects easily. It's very clear that the Red Engine was made to make games like Witcher and it's really good at it but that laser focus make it difficult to expand her to other genres just like Frostbite (for example the police system at launch was just repurposed from the Witcher 3). Worse yet if you have a goal of making sandbox games with living breading believable worlds, were you need a engine that does everything (shooting, driving, melee, flight, dynamic AI, realistic physics, etc) like Rockstar games. CD Project doesn't have yet the resources of Rockstar, that has a global network of studios and one focused mainly on the engine (Rockstar San Diego) to work on just one game for how long their want. As the Matrix demo showed Unreal 5 make it very easy out of the box to create such open worlds. They just have to customize it to their liking. They probably made the calculations of the work, money and time to make the Red Engine to become something like the Rockstar Game Engine and realized that it would be cheaper and easier to use unreal to get it there. Meta human alone will save them millions in creating believable high fidelity NPC for their cities, something that is a weakness of the Red Engine. And with unreal you get the added bonus of easy of hiring in every major development hub on the planet, you don't need to train personal.
Hopefully their engineers will help epic otimize unreal to be better at multi-threading. They will have years to do that before the game is ready to launch and as the engine is feature completed they can start working right now as the other part of the team do pre-production and prototyping and the engine team have little work on their plate.
DICE and other EA studios went through the exact same pains when trying to make the Frostbite engine the "EA engine" for all of their titles, while it was originally only designed for Battlefield games.
I think that EA and DICE managed it well and now Frostbite is a proper cross-studio engine. But there were many, many issues early on that had to be ironed out and Bioware talked about this when working on Dragon Age Inquisition.
So it's easier to just go with UE5 than trying to transform their single-studio in-house engine into a standardized multi-studio, multi-game engine. That would also mean having an entire dedicated studio just for maintaining and developing the engine itself.
The switch to Unreal engine has been talked about by the ceo, CDPR will actively develop on the UE too. This isn't a regular licensing agreement Epic does. So the manpower and knowledge of CDPR could very well improve the optimization for Unreal engine.
@vitorHugoOliveiraSousa BREAD :)
Was thinking the same the other day while playing Cyberpunk. The game looks incredible, I understand not having a specialized team to develop the engine can be positive in the long-run, but it's such a shame to abandon such a great engine. They should at least make it open source or something so it doesn't just die.
that game still has so many annoying graphics bugs and a raytracing implementation far less baked than northlight engine. The game is very messy technically, even though it looks beautiful, so I understand why they want a cleaner tech base and accelerate development that way
I think the big reason for the change is multiplayer. CDPR tried to implement some sort of multiplayer in cyberpunk, but they could never make it work. Unreal engine has a really robust multiplayer foundation to build upon.
There are also things that matter to game developers separate from what matters to gamers. I remember there being talk about how various issues with red engine caused problems during development. I don’t know what their tools look like, but it seems very likely to me that unreal engine has better tools for artists, game designers, and maybe engine programmers as well.
Hopefully they can work on the main thread dominance of UE5, though that’s typically really hard to retrofit.
does UE5 have modding capability's like the red engine?
because that would suck if it didn't have any modding capability, and would likely cause people to mostly stay playing on the old cyberpunk if this were the case.
@@user-tp5yb4hr4w Unreal can read loose files by default, and it can load scripts and data via DLC-like mods. But it's the developer's choice. Mass Effect games are extremely moddable as BioWare allows DLC slipstream as long as Table of Contents is recalculated.
@@DanielClear2
so your saying if the dev chooses not to allow their game to be modded, nobody can actually mod their game regardless the tools used?
i guess i never really considered this as a concept.
i suppose if CDPR don't allow modding of their game, that could cause many not to make the move over, once you allow modding, people typically don't want to use anything else if they like the modding capability more than what the new game offers in comparison.
i mean starfield is moddable, but people prefer to still mod skyrim instead, but that's partially to do with people just like elder scrolls more than any of the more modern day setting games as much, i could see people still modding starfield and it one day becoming popular, maybe not right now though, i think bethesda focused too much on bad game design for this game which makes it hard or not as easy for modders to want to mod given the design they chose for the game.
@@user-tp5yb4hr4w CDPR has worked closely with modders before, so I doubt they'll not choose to let people mod whatever games they make.
If CDPR had moved to something like Frostbite for example, then I'd be concerned, both with the game itself and the modding scene.
@@user-tp5yb4hr4w bro you dont know what ur talking about. Creation Engine 2 tools for modding haven't been added for starfield. These tools are already available for previous BGS games. Starfield is also a month old lmao. "LOOK AT ALL THE MODS ON SKYRIM AND LOOK AT ALL THE MODS ON STARFIELD RN. PEOPLE MUST HATE MODDING ON STARFIELD." actual brain dmg
It feels like a shame to switch now, even if I understand their initial reasoning. If you had asked me years ago when the game came out if the idea was a good one, I'd say yes 100% as the game still felt like a mess of spaghetti code. But now after the release of Phantom Liberty, it feels as though the engine is infinitely better optimized for PC's than anything that's released on UE5 in recent years.
We don’t know what kind of Frankenstein code and duck tape is being used to make the game work. There has to be a reason for the move, they wouldn’t be doing this on a whim.
@GD2X it's nothing to do with them doing it on a whim. They took the time and initiative to develop thier own technology through an engine that probably cost them at least tens of millions of dollars, for one game?? No, they switched for reasons beyond faulty code.Bad Publicity feels like it was at the core of the decision at least to me. I think they intended to enter the market with that engine and outsource it to other games . And become a competitor to Unreal Engine. But their showcase that was supposed to be cyberpunk was not given enough time and therefore they couldn't do it. That game needed 2 more years under the hood. The leadership lost out on 10 million plus sales I bet if they had just Waited . Instead of insisting of launching on last Gen hardware. And not allowing the base hardware for the other consoles to grow thier player bases.
@@GD2Xeven if it's duck tape is miles better than whatwver code unreal engine uses on PC, I actively avoid everygame that has unreal engine now, shader compilation is atroutious
@@destebangm11216 and it's usually fault of devs, not the engine.
@@destebangm11216 yeah, i don't understand why gamers worry about engines of all things!? unreal since UE4 has been unoptimized mess, and always needed high end hardware to even reach 60fps, current UE5 titles can't even reach 60FPS with using lowest quality of DLSS or FSR 2 on high end hardware, just look at Remnant 2, IOA or even fortnite at highest settings..
Didn't Epic say, when they announced their cooperation with CDPR, that the deal included help with building the engine? I understood that they will re-use whatever CDPR will build and ship it with future iterations of the engine. And it's not just CDPR, but other bigger devs, like Coalition. That's what I actually like about the deals they are making.
Yea something like that. It seems like they're using UE as an open pool of resources that are collected transversely from multiple parties chipping in - and in the end it means each company can focus more on developing their products, rather than developing their engine. They've probably also come to a point where they realize that true differentiation comes from scope of project, storytelling, and creative use of tools and game mechanics, rather than simply having better tools than the competition. The competition is already plenty capable of fucking up regardless of what tools they use, so this is really not where the battle needs to be had. So there really isn't any reason to gatekeep certain tools while handicapping oneself at the same time by trying to do everything inhouse. Just means that the actual differentiating aspects that truly matter, get less potential to shine in the end.
When Nvidia already works with the UE5 toolset and is actively developing it for their software implementations, and CDPR can get some good win-win deals out of it too, I bet they see much more profit to be had from helping Epic develop the UE5 engine to its peak potential, rather than continuing to work with the RED engine that is a development nightmare in terms of product pipeline for projects - as well as any new hire having to be taught the RED engine fundamentals, whereas Unreal 5 is open source and widely available, meaning when new devs are hired, it's more reliable to expect them to have proficiency in the engine being used already. This has always been the Achilles heel of any company with an in-house engine, when their senior devs leave or they have to expand their dev team, each new member has to be taught the basics from scratch. EA and Dice experienced this exact issue with the Frostbite engine for Battlefield 2042, where a lot of senior Dice devs had left and all the new hires had zero experience working with Frostbite, and thus the game was severely undercooked and shallow in scope compared to prior releases.
They probably moved to the UE5 engine for a few reasons.
Studios nowadays commonly use external contracted developers to save on costs. Hiring developers to use the Red engine means they waste a lot of time training them to use the engine which can be costly.
Right that's definitely part of it
UE5.4 road map has announced parallelization of the draw thread as a milestone. So I'd say that part is at least on Epic's road map.
UE5?? Get ready to buy Nvidia RTX 6090.
LMFAO
probably 7070 or 8070 hahaha
Nah, unreal engine 5 IS pretty well optimized with the "nantite" technology (i think thats the name) the only reason to buy an insane GPU IS if u wanna play 4K or 1440p with path tracing, and everything ultra, but u wont need It to be able to just run the game
@@cs1645 ue5 is famously poor performing. Both nanite and lumen are very costly to run on current technology. They build these technologies for the next generation of hardware, with scalability in mind. So they run pretty poorly on full scale games on current hardware. I have a 6800xt, which is still a proper beast, but struggles with ue5, unless I disable nanite and lumen.
I've been a professional Unreal Engine developer for 14 years. I've also questioned the move to UE5 as it's going to cause issues that the game doesn't have right now. Lumen is amazing but the development time to get correct lighting in certain situations can be strenuous. It looks nice out of the box but when you try style the lighting in a cyberpunk environment you end up with several performance issues. RedEngine has done really well to make CP2077 look as good as it does. However, the rest of the tech in the engine is absolutely beyond worthwhile for a new CP game and it's extended development. If they can overcome some of the lighting workflow performance issues, then they will absolutely get a better development cycle for future games.
As far as threading goes, the game thread is on a single thread but you can customise it for multi-threaded, even though it's a lot of work. Path tracing isn't important for anything except performance at the optimised RTX end, and it does well without it, but what path tracing does give is some extra effects such as caustics which can be faked any other way anyhow.
They managed to create entire engine that could rival other top engines, Im sure they will be able to fix thing or two that is subpar in UE when it's needed.
@@lorkano Absolutely. I've been super impressed with how they got CP2077 performance down on the PC. If they made the equivalent quality in Unreal Engine it would cost them a lot more in milliseconds by default and would require some heavy trickery to produce the same quality. Not impossible, but will take time. The benefits will definitely outweigh the on-boarding negatives though. Especially on the multiplayer side of things. Replication makes multiplayer really easy for them.
Have to wonder if the move was developer led or business led. Business people can get blined by the theoretical savings of some fancy piece of software, but struggle to see the actual cost. I see a lot of mid-sized studios go for UE and I think it makes a lot of sense for them because they will have some graphical vision, but they're more willing to compromise on it because there are fewer expectations and they just want to get their vision made at all. CDPR has a reputation and if they can't get the same fidelity that will be a huge blow to them. People will compare the next game to CP2077 so it better look just as good or better.
Any views on why Immortals of Aveum, the most recent UE5 game, ran so poorly? It has terrible graphical stuttering even in higher end cards. And while it looks ok, it’s not close to Cyberpunk 2077, despite being a linear game and a very short one as well (15-20 hours campaign).
@@pressrepeat2000 10 FPS in Matrix demo with RTX3080. That's why. I guess next-gen means less optimization now. I hate to be the hater but I cannot can't keep myself from hating when overmarketing and bad results meet.
"UE has a significant development advantage compared to RED. It's much easier to prototype and interpret game features. Visual scripting, in particular, is one of the most advanced aspects of UE in the world. You can provide the built-in tools to anyone, and they can quickly produce something. Try doing this with RED, and you'll need at least one programmer attached to the prototype.
Also, don't forget about the project-specific requirements. For Cyberpynk, they completely reworked the rendering and lighting compared to the RED Engine version of Witcher 3. Also some features, like the terrain and landscape editor, are not as well-developed Cyberpunk 2077. The Witcher requires the rendering of a huge, open world with lots of foliage, natural elements such as caves and bodies of water, and it's also a third-person game. Probably, CDPR made the business decision to move to UE because it's more expensive to develop these features from the ground up.
Absolutely, people should not forget that UE is not just a rendering tool, but a compositing one aswell. If not the best.
As a programmer I find visual scripting unnecessary. To be honest, anyone who knows what they are doing would find any visual scripting for any program unnecessary, limiting, and non-intuitive. As for the developers in CDPR; I assume they already got their hands dirty with raw code already, so I don't see how "visual scripting" (which is actually a frontend for not to threaten newcomers) is an advantage at all. The only explanation is that CDPR will be able to fork the UE5 project and make it meet their demands on-the-go, otherwise it does not make any sense at all to think of a triple-A video game developer to use a generic game engine. There is a reason that all gigantic games use their own engines.
Here are some examples of "good" large games that used unreal (in any version): Mass effect series (can be considered almost AA in current times), Hogwarts Legacy (which is more demanding with ray tracing all enabled than cyberpunk with path tracing, and stutter, oh the stutter), arkham games (the last one was in a bad shape initally, I am lucky to have played it in 2021) callisto protocol (which was even worse than cyberpunk at release, so does UE really help with streamlined game production I wonder?), bloodlines 2 (I believe still in development hell, can't say if it is about programming or creative direction though). What I want to say is that UE5 is not a magical solution for anything related to game development. It was greatly marketed and uses some advanced and clever techniques which are still not tested well enough by the consumers (due to lack of good games that use ue5 being released).
Here are some examples of "good" large games that used custom engines: All GTAs and RDRs using RAGE. All Bethesda RPGs using Creation Engine. All Valve games and many community-made games that use Source Engine. Death Stranding and Horizon Zero series using Decima engine. All AC games using various versions of Anvil game engine. And of course RED Engine with witcher and cyberpunk 2077.
Disclaimer: I am not a game developer, I am a backend developer for mainly web. There are some correlations between them, but one giant difference according to my one-year experience in Unity and UE. When we use publicly available frameworks (that are kind of similar to engines) we are greatly limited and always require people to be good at that specific frameworks. That's why in any large project; we tend to use a compilation of libraries (streamlined implementations of available technologies; such as being great at implementing directx in a game) that in turn form a framework speciel for us instead of large frameworks (ue and unreal).
@@echorises As a programmer myself who once made a calculator in my youth, I concur.
The main issue is with huge rotation of employees and time spend on training and this is the main argument for UE
@@echorises Wow? you want to program everything for the artists and game designers on your team when you can just provide them with custom nodes.
Something we have to keep in mind is that they had so many problems with the red engine I'm sure that way they can concentrate in development.
I remember how blown away I was with Batman Arkham Knight's lighting and overall visuals back in 2015. That game was build in Unreal Engine 3.
No doubt CDPR will amplify UE5 to insanity.
Arkham Knight still looks fantastic. Unfortunately I'm on Xbox, and even on the X it's, what, 900/30? That's pitiful. I was so disappointed when I tried the BC version, expecting 60fps, auto-HDR and upres'd, but no.
Such a shame. The bat-tank/Batmobile in particular has superb manoeuvrability and animation transitions, that's just crying out for a better framerate, at the very least. The fluidity of the gameplay is so damn impressive.
That's not on the engine, of course (it's on Warner, I assume), but still...
Absent mindedly playing with Unreal 5 tools and then going back to dutifully patching RED Engine probably feels like cleaning out a garage. You can either stay on top of engine development and tools thereof since the 90s, or fall behind until it feels like constantly reinventing glass.
From a business standpoint, it's probably better to pay the licensing fee on Unreal, broadly expand your hiring pool, and fix the hitching, than eat the downtime on an in-house engine. Now that Unreal has outgrown its FPS roots and has a proven track record with racing, adventure, and RPGs, there's as much to gain with Unreal for industry veterans as there is for someone who just showed up.
C'mon dude because you wasted far too much money on school that doesn't mean you need to flex your over educated brain on here
Just say "UE5 will be better" smh
You're welcome
Man Dachejaegers footage looks so sharp, the eyes really pop. Great setup you must have
But Red Engine run extremely well for the fidelity it gives. UE5 games so far has been all unoptimized mess.
The issue is the development difficulty, the logic to move is around being able to build games in less time.
Example of UE5 games?
@@crozravenimmortals of aveum, remnant 2 and fortnite all run terrible
I don't think there is anyway to tell. There is a high probability even CDPR doesn't even know the full implications of it yet. Unreal will streamline plenty for them but with anything its not going to be all pros and no cons.
Just from a end user experience though, the Red Engine clears does some things better. I assume that will translate into helping Epic with Unreal because Unreal needs some work on its efficiency and CDPR's partnership is probably a win for Epic on that front.
The questions coming from these people suggest that they don't understand how the engines work at all.
@@pawprinting There has been a ton of that lately in the form of criticism toward different engines.
Thats on the assumption they can completely overhaul the threading of the engine
@@pawprinting when almost every single ue4 and 5 game has stutters and basic shit not working correctly and horrible fps needing upscaling to even be playable I think all the backlash is right. The engine is beautiful but even it has yet to top red engines look idc how "realistic" it looks path tracing smokes anything on that engine and that aesthetic just doesn't seem doable when every ue5 game looks like a shitty realistic version of a game instead of a game with a art style and realism elements
"CDPR doesn't even know the full implications of it yet", but youtube commenter already knows
lma
People forget it took 10+ years for Cyberpunk to get to here since it’s initial reveal. Plus we still haven’t gotten Witcher 4 despite W3 approaching it’s 10th year anniversary since the studio was focused on 2077.
The switch to UE5 is to vastly improve CDPR’s efficiency (without compromising on quality) and to allow them to work on multiple games at the same time. The results produced by the RED engine are phenomenal, it just take too long to get there.
You forget that first trailer from 2012 doesn't mean that they were making the game, just that they have announced new IP.
Red Engine 4 is completely new engine build from scratch AFTER Witcher 3 and it's expansions(2016 - 2017). Designers had to work on incomplete engine and tools in the same time as CP was developed on this new engine was in the making for 1-2 years and they had another 1-2 years time for actual game development before initial 16.04.2020 release date. That's probably the biggest reason why CP was so buggy at release.
Now the engine is finally ready for production as we can see, but well, it is what it is.
And it's ever more insane if you think about that UE5 is just overlay for UE4 with nanite and lumen added, and after 20 years of development it's still behind what Red Engine 4 can do after 5-6 years.
@@Execell0 'UE5 is just overlay for UE4 with nanite and lumen added' - I kinda agree, but saying 'just' is insane. Cerny was talking about Ninite-like tech, when discussing PS5 originally, and still, 4 years later, only Epic has delivered it. That shows you how difficult it is to do. If CDPR could deliver something like that with Red Engine, that would be game over, the best engine ever. But they clearly could not. At least not in a time-efficient manner.
@@Execell0"just" only two of the biggest leaps in gaming technology of the last decade lmao. 😂😂
There is not 1 game using Unreal 4 or 5 that comes EVEN close to the feel and look of Cyberpunk 2077. Like its not even remotely close.. Unreal has always had a shit look and before you say its Art Direction sure, partly however every Engine has a look and feel that is almost ALWAYS associated with that engine. If this where not true then every engine would look the same but they dont. They all have different ways of doing the same things which results in differences. Sure they might be able to develop games fast but unless the heavily modify the engine this will surely be a downgrade in the future unfortunately.
UE5 has ways to fix traversal stutter. It just takes work. You can group objects to load how you want. But that takes time to categorize everything into groups and when to load them. So you load big things first and slowly bring in the small things as you get closer. But that takes lots of time and money for performance most gamers don't care or notice
One thing that must be taken into account is that the next Cyberpunk game running on UE5 won't happen for roughly 4 years. The "next-gen" consoles will be launching, and the engine itself will have matured considerably in the 4 years.
cyberpunk game will take way more than that, they will be working on new witcher on UE engine first.
lets hope the engine is optimized. i use it daily and its a shitshow
they will be working on 2 project in one time. But yea, Witcher is 2/3yers in production so cyberpunk will relase 2-3 years after Witcher@@lorkano
Four years GPUs will be improved.
@@lorkano Orion(Cyberpunk 2) and Polaris(New Witcher mainline series) are being worked on by separate teams, with Orion being made in their new Boston/Vancouver studio and Polaris at their Polish one. Then you have Canis Majoris, a side game set in the Witcher universe which is being worked on by a studio that probably ported one of the previous Witcher games, and Sirius, being worked on by The Molasses Flood which CDPR bought out. Then you have Project Hadar, which is an entirely new IP created by CDPR still in the groundwork stage for fleshing out the IP itself, without any actual game development started.
Shader compilation is also another big concern with the move to this engine. It’s become too big of an issue lately. It’s a shame they’re not sticking with their own engine. They’ve put so much work into it and it looks amazing. Every new AA/AAA game seems to be using Unreal engine now and they all feel the same.
Not all UE5 games have shader or stuttering problems.
What are the big AAA games that use UE5 ? All of the games I've seen using it are like Lords of the fallen which is way smaller budget games than something huge like Witcher or Cyberpunk
"Every new AA/AAA game seems to be using Unreal engine now and they all feel the same."
While there are a lot of game being built on the engine hardly any UE5 games have actually released......
He didnt say ue5, he said UE in general, and ue4 games doesnt have a 'similarish' feel to them@@t3chfx13
@mordor1779 Immortals of Aveum.
Its good cuz they can focus on story, art, music. Also UE has good mp foundation, also nanites will make huuuge impact for game in case of cities which can be bigger and denser.
No it won't. There's a reason no UE5 games make use of nanite much, other than in small segments/areas, and there's a reason why there are no large scale open-world games made in UE5. Nanite is not a magic bullet how people think -- it comes at a cost to the resource budget as the play space scales up. The multiplayer is a good a addition, but I feel CDPR is losing A LOT of optimisation real estate moving from the Red Engine to UE5. It would take a bare minimum of two years to refactor UE5's low level code to get thread optimisation on par to Red Engine, otherwise they would just have to cut corners and we would see the same sort of frame hitching and processing issues we see in the Cyberpunk sequel that we see in every other UE5 game.
@@billywashere6965CDPR will customize UE from ground up to their picture. They will use experience and make it work, trust me. And in terms of large world, just wait that's all i can say.
This is a great question and one that's been on my mind for a while... I think the Red Engine is fantastic but I wonder what the Unreal engine can do to really upgrade the Cyberpunk sequel in the works. Thanks for covering it, I'm not very technical but great video explaining it.
The move to UE will bite the industry in the ass sometime.
Agreed wholeheartedly with you. Just a bunch of same feeling and playing 'games' with similar technical issues that in some cases never get resolved.
@@99Vood99did you forget how ps2 era was? It was all renderware games lmao
@@henrik1743 Yup, history repeating itself.
Incredible video. This is what I was looking for, people who know their sh-t. Thanks for the very educational take on this big question.
You choking on those sausages?
What's that blue bird icon in the lower left corner of each frame?
I'm pretty sure the other day they said they are working with Epic to port Red Engine features to UE5.
Cyberpunk on my 1050ti is shockingly smooth for my settings, so happy woth how ot looks its a step above anything else
Before I got a new system with a 3080 I was running Cyberpunk on a 1070 and was shocked at how well it ran.
A major reason they are making the switch is most likely due to labor costs and availability. They are opening a US office, but around the world there are people experienced in Unreal that require much less cost to train or to replace compared to people skilled in their own engine. If they lose talented (and thus expensive) programmers for their own engine they are not going to easily replace that skill and knowledge cheaply.
With how poorly they treat their employees in terms of hours and layoffs, I’m sure that this is a major factor in their decision making as it allows them a much wider group of employees to burn through, and less cost when a high end employee leaves.
I imagine their employees are aware of this as well, which was probably a factor as to their unionization.
In short, it’s not about the power of the engines, but labor logistics and $$$.
Which is a shame, because UE5 is no where near as optimised as the Red Engine on a low level; Red reads to the metal so much better than any UE5 game, especially with Cyberpunk being open-world. If they don't refactor the UE5 for low level optimisation, then Cyberpunk 2077's sequel will run worse than how Phantom Liberty runs right now.
@@billywashere6965 im suprised that epic hasn't fixed this yet
fun fact, when you join a AAA company that uses a in house editor you don't get trained, you have between 2 weeks to 1 month, maybe even two to get up to speed with it and read internal documentation
correct
@@kaos88888888 That's 2-weeks to 1 month of paying someone just to learn though, and then who knows how many months until they actually reach the knowledge and output of a veteran developer. It's much easier and cheaper to hire someone who is already familiar with the engine you are using, and also takes away a LOT of leverage from your own high end employees that are experienced in it when it comes to raises etc. when they know they can be easily replaced by a huge number of people that are already familiar with the engine the company uses.
Cyberpunk, especially the stuff in Phantom Liberty, looks absolutely amazing.
"If you move around in any world in UE5 it's stutter city".
So far, this was true for me in Ghost Wire: Tokyo, Eximius: Seize The Frontline, and Robocop: Rouge City demo.
Maybe, just maybe it's extremely expensive to continuously develop inhouse engines every time a new project is under production. And we also need to weigh in the fact that people know their way around Unreal engines so hiring new people will enable them to start working right away instead of being having to learn a completely new engine they've never seen before.
And maybe, just maybe CDPR also realises RED Engine has specific limitations of what it can do, regardless of having awesome frame latencies. :)
could be that, but i dont see the problem at least from the end user perspective of this "limitation" on the game, 2.0 is the proof that the game can run flawlessly and look amazing
its not like they making a whole other type of game like Mass Effect Andromeda and Frostbite situation when they basically making all the game code and features from the ground up for that engine who specifically created to make Battlefield games
CDPR future titles is still in the same game genre
Which is sad for us consumers because that same thing happened the FOX engine which had only a few games on it and the graphics capability and optimization of that engine blew everything away back in the 2015, and I see the same with the Cyberpunk and RED Engine, the graphics of that game is out of this world and the way UE renders make every game made with it pretty similar to be honest, which honestly is sad.
Except unreal 5 is awful.
@@diogofelix8626 CDPR has done work with RED Engine for 2.0 and PL. UE5 or UE-whatever-it-will-be, will have source code available, just as the gentleman in the video said. This means that graphical stuff can be imported or rebuilt. The problem still remains that UE sucks on single thread performance, but this aspect can only get better and better. Sequel for Cyberpunk 2077 will take years, half a decade or even closer to decade, so we can't really make any predictions about the performance or maximum graphical quality yet. New generation of consoles will also hit the market before sequel is done, and it's very likely that those will be supported.
Bethesda should take your advice lmao
Has the UE5 engine proven itself for open world games? Does the UE5 Matrix reloaded demo have CPU issues or performance issues?
Perhaps rumors the Matrix demo ran on Switch 2 (doubt actual silicon but ballpark target hardware?) maybe shows improvements can be made towards the engine's cpu issues...
not gonna happen, wait for next gen CPUs, nanite is very single core CPU heavy(from their GDC presentation) and still doesn't support direct storage or other dx12 ultimate optimization features..
as unreal developer ue5 is still not production ready for open world game design due to landscape system issues, virtual texture issues/virtual heightfield mesh issues, no tessellation for landscapes, various other graphics/rendering bugs, it's not just a good every-use-case engine - maybe if you have whole team of programmers that are fixing engine issues to match your game design but overall at the very end, as a developer, you end up with a barely working product, most of developers are still sticking with ue4 or even ue3 with reworked graphics code like MK11
UE 5 Matrix demo have no performance issues if you run it from M2 SSD. S-ATA III SSD makes it stutter as it can only read 550.MB/s.
HDD? Forget it. Stutterfest.
@@niks660097 Darn thats unfortunate, as improvements in IPC and clockspeeds for CPUs have slowed to a crawl though, and even worse CDPR's next game might be developed with the CPU limitations of current gen consoles in mind with their ~3ghz half cache Zen2 cores.
@@V3ntilator also remider that it doesnt have any advanced ai or missions or anything else. even then it stutters
From a business perspective it must make sense I guess, keep engines up-to-date cost a lot of resources and it will take you longer between releases. My concerns with ue5 is that the overal look and feel, especially with third person games will be the same no matter which studio makes a game. That Studios and IPs will loose its identity in said way
CDPR is limited by its engine that does not support multiplayer in any form.
And making it would be a nightmare, a total rework because it never supposed that as a possibility.
Also the lack of scaling for larger teams is a concern.
To make UE work well they can have support from Epic, like they have now support from nVidia.
CDPR said they will be implementing features from red engine into UE5 and taking the best of each while working with epic to make improvements. I honestly think this will make UE5 improve substantially.
CD Project switching to UE5 must mean they hit a ceiling with their current engine, and switching will be more cost efficient in the long run.
It's hard to speculate what that is exactly, whether it be a price ceiling or a technical ceiling. Maybe porting the game to other systems was the culprit?
I'm more optimistic about this move since their next game is years out, giving them a good amount of time to switch to the new engine. I can't wait for their next game release.
Finally. Someone who understands game engines or at the very least, business lol. People think the red engine is good, It isn't and even CDPR admits that. Kind of the whole reason they want to use UE5 for the sequel lol.
yeah I'm not worried, they got the technical experience to highly customize it to suite their games for whatever they need. They just did a presentation and they're already modifying it
CDPR has literally said that red engine is at the end of it's lifespan and they must move on.
The only thing I'm worried about is shader compilation stutter like we saw in Callisto Protocol and other Unreal releases. Part of what makes Cyberpunk 2077 so great is there's minimal stuttering and it's shockingly well optimized
Or they fired / lost the core team working on the engine already.
It's about scalability mainly, and they will add what was great about RED and integrate it in the UE5 engine.
The fact that I can run path tracing with an RTX 3070 in many areas of Cyberpunk 2077 (albeit using performance DLSS) is insane to me. It's crazy how well optimized the game is. After experiencing problems with Callisto Protocol (a UE4 release) and extreme shader compilation issues I'm definitely worried how performance will be. I've noticed unreal engine games usually launch with issues
I'm excited for how CDPR can use things like Nanite and Niagara, but yeah worried about performance
runs very poorly on AMD, so I'd say its not well optimized.
It is nowhere near well optimized game. People forget it is a ps4 era game. "Performance DLSS".
@@theriddickit runs very well on my AMD Rig
@@76marex 7900xtx, 4k, 50-60fps without upscaling, 60-70 with. Upscaling both XeSS and FSR2 introduce considerable image quality reductions. XeSS seems the best for clarity but does have artifacts still.
Enabling RT gives too low a frame rate, and PT kills the frame rate.
Well, CP 2077 did launch with issues as well. It took quite some time and a lot of work to fix it and make it the game it is today.
Any one really believe that UE5 being "a work in progress" will change anything for CPU utilization?
This is a core problem, the most basic thing, the one thing you expect it to be fixed first had no improvement.
No way that CPU utilization will ever become "good" on UE5.
It's only gotten worse over the years, I have zero hope it will improve.
@@stevy2"Zero hope" if the same group of people kept working on it. But now CDPR's devs will contribute. Whether it's their own branch or the NVidia branch or something else, you're going to have the same people responsible for RedEngine working on it. That bodes well for UE5.
I guarantee they're not making this transition without have done the homework and balanced the pros/cons. They know their own engine intimately and they've obviously looked at UE5 and done the math on what they need to change for it to serve their needs. Any issue with UE5 is going to be something they've looked hard at and said, "We can adjust that to fit our needs."
The part of the calculus that few are talking about is the friction of RedEngine. They wouldn't be making this switch if UE5 didn't offer them something that RedEngine did not. Or, if adding new features to RedEngine was going to be more laborious and time consuming that simply using UE5 or adding those same features to UE5.
This change is about speeding up their pipeline, removing friction/pain points, and removing technical debt that has accrued from years of building their own engine. Considering they wrote the RedEngine, I would trust their process here.
saying it will never become good is a bit ridiculous. red engine shows that excellent multi threading can work, just needs someone to implement it in UE5. i wont hold my breath, cause it certainly wont be a quick and easy fix, but its certainly possible.
at the end of the day unreal engine is nothing but a collection of modules dedicated to a specific task. ue5 is ue4, just with 2 major new features added (yes and small incremental changes here and there). someone just needs to work on load balancing.
Its unfortunate because the RED engine is gorgeous and looks pretty unique, But I understand why they want to leave it behind. I'm just worried we're about to get a wave of AAA games that all look the same, because of UE's very noticeably rendering style. These companies aren't used to having to stylise their visuals for uniqueness and I think most of them will just ignore the process.
I'm more worried about needing a 7090Ti to get UE5 games run above 30 FPS... Sure there's a lot of eye candy in UE5 (but so does cyberpunk!) and these glaring optimization issues seem to be really difficult to iron out otherwise we would already see several UE5 games in large open world context.
@@Bourinos02 "these glaring optimization issues seem to be really difficult to iron out otherwise we would already see several UE5 games in large open world context."
UE5 came out in 2022, massive open world games take anywhere from 3-8 years to make. You really can't expect a true UE5 game to come out before late 2024. So far only games that have come out on UE5 have been imported from UE4 and are not true UE5 games.
@@teemumiettinen7250 The matrix awakens UE5 demo is evidence that UE5 isn't suited for this kind of games just yet and that a loads of workarounds and optimizations will need to take place for any game dev to use it for such purpose.
@@Bourinos02 way to say nothing but pointless off-topic crap about things that dont matter at all to the timeline discussion. You can't expect true UE5 games until late 2024.
@@teemumiettinen7250 You said it yourself, some games are just being ported from UE4 to UE5, why aren't they?
I agree with the sentiment that most people expressed. The cost of maintaining an in-house AAA game engine has gotten so high/expensive (labor, dollars,time, etc) that outside of big AAA studios (EA, Ubisoft, Epic, etc) most mid-size studios can't bear the cost.
A highly underrated point is that if CDPR has multiplayer aspirations (which they clearly do), adopting the REDengine (a solely single player engine) to support that will be a massive undertaking. Even the most talented of single player studios (looking at you Naughty Dog) are facing almost insurmountable challenges (that Last of Us multiplayer game) in doing so. Using the UE5 engine (which supports arguably the most popular multiplayer game Fortnite) will greatly speed up that aspiration.
cant wait for stutterpunk and witcher a wild shaderloading
Unfortunately, we steer into a direction where way too many big AAA studios use Unreal (except Rockstar with RAGE), similar to how everyone used UE3 back in the 360/PS3 era (we still had Engines like CryEngine back then though, which is now only used by Hunt Showdown and Sniper Ghost Warrior 3 afaik).
Don't get me wrong, UE5 is great, but I think different genres of games have different needs and therefore require different engines. I don't think it's desirable for the gaming landscape as a whole that we move towards what is essentially a UE monopoly.
Take Bethesda games for example: the Creation engine may be old and require way too many loading screens, but the item persistence it offers is not currently possible with UE5 (correct me if I'm wrong).
I just fear that everything will revolve around the best possible graphics and textures while gameplay related advancements are put aside. For example, I would love to see improvements in physics - it seems like the Euphoria Ragdoll physics in RAGE were the last major advancement in that regard.
If anything, it would be nice if bigger companies were to use (and build up/fund) Godot Engine since that has an open licence.
Why would you think that would be the case? If anything, if they’re using Unreal, Epic Games would be doing a lot of the heavy lifting and R&D on the graphical end, which frees up more of the programmers to work directly on gameplay-related programming, instead of also having to reinvent the wheel all the time.
Additonally it make each game look LESS unique. Engines have their own unique look.
Creation Engine 1 is old, but BGS got a lot of mileage out of it before moving up to Creation Engine 2, which is what StarField is built on and it's what TES6 and Fallout 5 will be built on.
Skyrim, Fallout 4, Skyrim SE, and Fallout 76 were all built on Creation Engine.
Creation Engine does what BGS needs it to do, plus it's really modder friendly.
@@JonathanJuanA single company without competent working on a single piece of tech won‘t bring as much innovation as a naturally competitive industry
It's amazing how Lies of P uses Unreal Engine 4 and was almost perfect if not perfect on launch
It also has nowhere near the amount of AI on the screen, the density and geometrical complexity of Cyberpunk and doesn't look as good in terms of pure rendering. A much more manageable and straightforward scope.
I think Atomic Heart uses UE4 too and It runs flawlessly. The devs did a great job.
Lies of P and Atomic Heart are not RPGs, so they are nowhere close to handle such logical complexity, as we see in Cyberpunk. Not to put down the effort those devs made to improve it, but it's a completely different level of complexity.
@@Heruwath007 Just take the example of Callisto Protocol or Jedi Survivor and the traversal stutters and the animation skips
@@JohnJohn-pm9wq And what for should I use those examples?
That makes me interested. What Engine will they use to create the Witcher 1 remake? Does Fool's Theory have their own engine?
How will the modding be with unreal engine? Will it be a locked pretty box?
CDPR is a big company; they certainly did a cost analysis of how many devs it takes to maintain and update the RedEngine vs UE5 and Epic taking their cut.
I don't think this an artistic decision by the devs but by upper management, they can now lay off engine devs and claim no more crunch because less work.
Yep, this decision has management written all over it, and it's likely going to bite them in the ass in the long run.
There's a human factor being ignored by many: some people really want to work on proprietary engines. Usually those people are really hardcore and very rare. Those are exactly the people you want to retain. Instead they'll move on, or be fired like you said.
@@MacGuyver85 i mean CDPR just laid off 100 people in the summer, lets be honest those coudve been from the engine department
CDPR is taking a huge risk with this as in a few years they will not be able to switch back and will be completely dependent on Epic. And most importantly on Epic's pricing which can be anything Epic wants at that point. Epic just announced they will start asking per seat prices from non-game developers and nothing stopping them doing this in the future for game devs as well and UE5 has no competition on the market for AAA devs. Monopolies like that usually don't turn out to be good for their customers....
They can modify any component with their own code. They will not be “completely dependent”
CDPR will have its own custom license with Epic. Also, Epic starting charging non-game developers is perfectly understandable and in fact expected change.
You are right but they also have a special deal with Epic
@@irecordwithaphone1856 The deal will expire sooner or later and Epic will have the upper hand in the next negotiation."
They'll just pass the cost on. And you'll still buy it. It's not a problem.
Wait until they are able to do quick asset flips to completely different game types, and we get a Night City Racing Game. The spin off games could be hardcore with how easily these worlds could be used for differing game types. Imagine a Night City Splinter Cell-like game. The other thing is the upcoming shift to cloud-hybrid games, will greatly expand the world capacity. You could create a Night City with entirely fleshed out districts the size of the current Night City. Then build it over time, even update it over time with refreshed graphics. Doing all of this without a single patch to the game, because the code updates in real-time across all players. The future is going to be cool.
It would be cool but also even more work and games will become even more expensive to create, if developers are expected to create that same night city depth into multiple night city sized districts. I feel like people's expectations will become crazy
@@irecordwithaphone1856 Well, it's built over time, the idea being that you make a "platform" to build a game on, and then just keep building on the same game over 10 years. Think Destiny 2, but with the technology to actual realize something with that kind of scope.
There are no "black box" parts of UE5 for game studios, we have full source and build it from scratch.
The real reason why they move on to UE5 is because most of Red Engineers left the company. Talents come and go and it is hard to maintain your own engine if the people who know the code inside out left the studio. It also costs a lot of money (also to train new people on the engine). Lots of developers know UE and thus for a big studio hiring new talent is easier and the tech is cheaper to maintain than with a proprietary engine.
bingo.
The people who left wanted to make red engine similar to rage engine I feel like with unreal their gameplay will become even worse
If they were developing their own engine they should have also planned to market it to other developers. I have always found it weird that they did not. They developed this massive, expensive tool only for themselves to make money off this tool indirectly by sales of the game instead of also licensing the engine.
From experience I can tell you that the engine used in house is rather different from what you are used to see in Unreal or Unity. Most of the time there is not even an editor or tools and if there are such tools developed then those are specifically tailored to the game that's being made. To fully be able to license it to others, the tools must be bulletproof and needs lots of features that other engine have out of the box so that a broad audience of devs can use them. @@rokhamler3352
@@rokhamler3352 the issue there is that having an engine fit for purpose for a single product - vs having an engine in a state you can sell as a product/service in and of itself for other devs to use for other products are very different propositions.
Just for one example, on launch, Cyberpunk had loaaaads of redundant duplication of processing in shaders. It was a crapshoot bloated nightmare that's taken them the better part of the past 3 years to iron out, it's in a great state now, for Cyberpunk to run well & look good, but that doesn't mean their tools are accessible or useful for 3rd party devs, it doesn't mean it's well suited to other applications, environments, art styles, gameplay styles .etc
It's like imagine if instead of having the Photoshop UI you had a disparate collection of different batch files & py scripts which need to have parameters & variables manually input every time you wanted to use a tool or effect or something. Even if it can produce the same quality output in theory - you can't market that as a product.
Remember Cyberpunk also had huge periods of crunch for workers - this might have been a culture thing but it can also be in regards to the workflow of the engine they're working with. Perhaps there are less superficial things about the workflow aspect of UE5 which will help them to deliver more efficiently and consistently and better for the people working there
Also, cuts down training. UE5 is an engine most developers know how to work with where as Red Tools, they would often have to slow down development just to train new hirers on how to use it. This was something cited as to why it took so long for Cyberpunk to release, was because so many members of the staff didn't knew how Red Engine worked.
Lol. You paint turds gold in your spare time dont you.
I don't think this is related to engine that is used. If they start making AA games instead of AAA, well yeah UE would help with that. If you are aiming for greatness; UE5 out of the box would not really enable anyone to do that; on contrary, it will be required to advance the engine if they don't want to have generic game. Do you know what is more difficult then polishing your game engine? It is polishing someone else's engine.
It is true that it will cut down the training period though.
Technicals well yes but what about RPG design ? RED engine was designed to create RPG games. There are systems and tools in RE to help you with creating multiple dialog options and entire quests for RPG game and this is something that CDPR games were best in - RPG design and quests. Hope they can bring it along.
Hey guys, could you do a tech breakdown of The Talos Principle 2? It runs on UE5
They moved to UE5 to cot down on their expenses. 150 people got sacked several months ago, a lot whom where people who were maintaining and working with red engine. My mate was one of them. Long term might be not best move thought.
its a shame what short term profit motivations has done to the gaming industry
It would feel like a shame now that they've fixed up the game quite a bit, the engine clearly does work. Even if it's a very resource heavy engine. I can still see the merit of the switch for optimization purposes. But it does seem like it'd be wasted dev time now. I don't think it's confirmed that the next Cyberpunk will be in unreal either, I thought that was just the witcher remake, perhaps to test the waters. Hopefully if the extra development has an impact on the games, they're able to get things fleshed out with witcher 4, so Cyberpunk can get the clean launch it deserves down the road.
They have confirmed all games going forward, including cyberpunk will be moving to unreal. The new director for Cyberpunk2 has talked about the transition.
@Hyde1415 That was a horrible decision by CDPR.
@@billywashere6965 we won't know until witcher 4 comes out I suppose
@@Hyde1415 Cool ig, hope it works out for them. I don't like epic, or the control it's trying to gain, but I respect the tech of the engine, and the documentation/feature accessibility of it. Truly is a good smooth product for devs to use unlike previous iterations of it. The engine for CDPR games also likely hit it's limits with Cyberpunk, and can only go further through Nvidia adding more complicated algorithms to raytracing to hack around computational limits. I really don't think it NEEDS to go further, but the transition gives them space to grow in that sense. As long as the releases are clean, shouldn't backfire on them.
@@MoistChickenLegs Yeah, I try to stay neutral when it comes to the business side of Epic, but there's a lot of good tech there to make the Cyberpunk everyone wanted. The biggest thing I hope they tackle, are NPC's that feel like they have life to them and better AI for NPC's on cars. Even in this update, we don't normally see NPC's on bikes unless there's a random shoot out event. This is cool, and definitely an improvement, but the game needs a bit more dynamic qualities for NPCs that I think the RED Engine struggled to do. 2.0 definitely did fix a lot of those elements though.
I’ve heard rumors that Hanger 13 is also considering UE5 for Mafia IV. It’s a shame because Mafia: Definitive is such a stellar looking game. They took that Mafia III engine and improved it so much. I’m also worried Mafia’s classic vehicle physics will take a hit as I have yet to play an open world Unreal game that had decent vehicle physics. I’m not well versed in game engines; could Hanger 13 essentially “port” their vehicle physics/coding from their current engine into UE5 and it would feel the same?
yeah you can basically "plug in" custom parts to unreal, so they absolutely can port their driving physics in there.
@@Kratoseum Alright that alleviates some concern then. Mafia has always had dope physics - it’s like a staple of the series, especially compared to other open world games that don’t focus on it. Every time I play GTA V I feel like I’m controlling little RC cars. Hah. So I’m hoping Mafia retains its more grounded physics based driving.
Decent engine engineers are expensive. Having familiar to everyone tech such as UE5 potentially increasing iteration predictability. It will also improve prototyping as you can potentially hire designers that are already familiar with editor.
Honestly I feel like it's a good move. From my perspective most if the bugs and issues that Cyberpunk had for almost 3 years were engine bugs. In simpler terms, it wasn't the code of the game but the code of the engine that they had to fix. Obviously not everything but a large amount of them had to have been engine issues straight up. Whereas if they use unreal they can focus the people who originally worked full time on fixing their engine can now focus solely on the game and optimization.
Exactly .... using a game engine to create a game and creating a game engine itself are two entirely different skill sets ..... Just making a DX11 game engine work with DX12 almost needs to be done from the ground up as both CDPR found out with Cyberpunk and Ubisoft found out with Valhalla (and Mirage too from the looks of their support Discord) ..... Moving forward both companies need a new modernized game engine written specifically for DX12U and not just some hacked and kludged together DX11 game engine converted to DX12
Trying to make a game engine designed for a high level API like DX11 work with a low level API like DX12 is almost a fool's errand ..... Threading is done mostly by the API in DX11 but is limited in ability while threading in DX12 has to be done by the game developer specifically but when done correctly produces outstanding results not possible with DX11. DX11 holds your hand while DX12 does not and takes extra effort.
Very positively impressed with the technical expertise of the team.
Honestly I dont think it would be worth it, but I guess the higherups have decided its easier to hire/scale up via ue5 and it has more tools/assets/plugins available that help out with things vs having to work all of it in-house and having to train new folk on the in-house engine too,
I still think red engine is an achievement and if nothing else they could do a free-license it out for people curious about it/modders, or opensource it
This. I wish they would do what CIG is doing and train uni students on their engine middleware and toolset workflow. Reinvest into their local game designers to work with the Red Engine. They finally got the thing to purr correctly and the Red Engine stands out far above and beyond anything else on the market right now, especially comparable UE4 and UE5 games, and yet they're abandoning it after they finally optimised it? Seems like such a waste.
The amount of work to deploy on multiple systems with an inhouse engine, without monetizing the development past your own game seems not to work anymore. Fun thing is image quality, and perceived quality reached top levels on previous gen titles specially those that were also on pc.
So Cyberpunk solves their frame rate issues but now goes for UE5 shader compilation stutters on PC??
They can solve that issue and so have other UE5 devs.
Traversal stutter is the real issue of Unreal Engine
@@JazerMedia why take the chance, they might not?
@@JazerMedia not a single dev has solved UE5 stutter. not even epic can solve the stutter in their new demo
My colleague at work was delighted, dare I say he's one of those Unreal engine simps who thinks it's the best game engine known to man and won't stop talking about it every minute of the day. When he heard about this it was like music to his ears. He sees any other game engine as inferior despite not knowing anything about other game engines or game development. He was also delighted when Unity raised their price and everyone complained. You actually can't argue with him about anything either 😂
Does he actually develop games in UE? Because I'd be hard pressed to find any dev that doesn't eventually grow to have a love/hate relationship with their engine
Sounds like a Lolcow
@@jmanners unironically no. He just loves it. There are no shortcomings 😂He made a few minute video for something at work which wasn't asked of him, he did it in his own time. He actually just quit the job to do a course in Unreal. He is in his mid to late 30's. I'd say fine if you 18+ or fresh out of Uni. The other irony is you can't get out of him what he intends to do after the course. It's like getting blood out of stone. My boss and I think it's career suicide, with my knowledge of the game industry it's like the fashion industry, it's a dog eat dog world so unless you have something unique to show for it you don't go anywhere. On the plus side at least I never have to hear about it anymore 😂
@@kevo300 yeah you could say that 😂
@@JohnDavidSullivan Anyone that has worked with UE5 will see it's shortcomings and bugs that are yet to be fixed by epic. If he's serious about doing development in UE5, he will soon notice, but he'll probably be running empty scenes anyways lol.
The fact that I don't see any shader compilation stutters on Red Engine, even though it doesn't pre-compile anything, makes me dread the shift to UE5 a little.
"Next Gen" update's tech side was done completely by Saber from what i've heard. Move to UE was mostly to be able to easily hire people to work on games, instead of having to train them to use RedEngine tools
UE5 makes sense for CD Project. Cost could be lower as they don't have to maintain their own engine (although they will have to pay 5% of their revenue to Epic), hiring and training new developers will be faster and cheaper as they don't need to learn a proprietary engine and maybe it will make their workflow more efficient with nanite
and since the pool of programmers already working with Unreal is massively bigger they can lower the wages
@@David_randomnumber No, they can't because good programmers would go elsewhere. Unlike devs that have been working for years on the Red Engine that would have to learn a new engine if they leave.
The 5% figure is only official number. The 5% may come from somewhere else anyway, like technological collaboration. Or advertisement. Deals between these kinds of big companies are never public about such licencing matters.
@@OverJumpRally But someone who already know the red engine is in a vastly better position to barter for their wage since all projects with this engine have a deadline and you can't hold this line while you train replacements. Whoever works with red engine and not already prepares for an alternative should maybe look for another job since all engines have a limited lifespan.
@@David_randomnumber yes and if you offer me 20K for UE5 dev, and this company offers 25K. Yeah I'm off
I think it's for the sake of reliability and bug squashing. Having just finished Phantom Liberty, the game is still quite buggy (both base game and the expansion), and a lot of the time the scripts just get stuck for no reason. I have had to reload the save many times because a charater would forget to move, or the game wouldn't give me full control after a cutscene.
If it still has these basic issues after years of fixes, that to me means they're basically unfixable, and while the game looks stunning, you can't have a 10% chance that something breaks at any given time. Basically the whole industry is collectively working on UE right now, there's support, and new hires will have already had experience with it. Easier, cheaper dev, more reliable engine. It homogonizes the landscape, sadly, but it's the practical choice.
There is just a lot of technical debt that they probably wouldn’t be able to shed without a complete engine overhaul which takes a lot of time and money.
It always blows my mind when people say the game is still a buggy mess and that they have to restart it. I have 500 hours in cyberpunk and haven’t had any issues the last 6 months.
@@braukwood925yeah that sounds like more of a hardware issue for OP than the game.
@@Tuhoeterra no, I can confirm game is still buggy. I am not able to complete the main storyline as im stuck trying to use terminal sierra delta. funnily, as was able to play through the main game right after launch without any major issue
@@braukwood925
I just played through the entire game and the DLC for the first time recently and never noticed any bugs other than some vehicle collisions acting funky when hit at high speeds. Certainly nothing game breaking.
Unreal engine 5 has so many advantages that its a nobrainer to work in. Making massive worlds by a click of a button once you have made a small themed area as a reference. and being able to rearrange and the engine adapting the surrounding by itself accordingly. and in the future not needing to bake in shadows, lighting and reflections. saving lots of time. the more studios who move to unreal the better. there will be more games and larger gameworlds and the standard for graphics will be much higher even from indie studios. with more games out, the prices gotta drop too.
I was excited about them moving to UE until I made the mistake of buying Jedi Survivor after patch 7.
The thing is, RedEngine is broken literally since its first release, and they are in constant loop of fixing/patching its issues, and is somewhat hard to work with, just look at how long it took CDPR just to fix Cyberpunk`s biggest issues, and yet there are still some issues present.
Even from mine mod creator`s point of view, a lot of RedEngine stuff don`t make sense.
By switching to Unreal Engine 5, they will develop games easier (faster) with easier way of fixing bugs and issues, and with all of the the tech (current and future) that (is/will be) supported out of the box in Unreal Engine, they don`t have to bother implementing them themselves, and all of the time they would have to spend on fixing and upgrading their engine further, they can actually spend on improving and polishing their new games.
Its just more convenient and easier overall engine to work with compared to RedEngine, and they can solely focus only on games from now on.
bro they literallty have to develop the engine to cater to their needs. epic isnt going to develop the engine with path tracing and openworld with ai optimization and large scale world. epic will improve it only for small devs and rest will be done by the game devs
I used to like unreal engine back in the day, now I absolutely hate it.
Skill issue
CP2077 now is such a great PC game. I have a 5700X and RX6800, and for the majority of the game i get around 100 fps on non-RT ultra settings at 1080p. And even when the CPU is going full tilt at 100% usage there's no frame time stutter, it's magic.
How? I get 80FPS on 1440p due to CPU bottleneck with my 7900XT when turning off RT
The Red Engine isn't flawless but the current build of CP77 is a remarkable achievement in rendering technology and I can't help but feel like its a waste to just kind of abandon that. I think it makes sense for onboarding their new employees, in their recent expansion, to have an engine people are familiar with so I get that aspect but idk I feel more dissapointed. Again, I'm not a developer so I dont know what I'm talking about when it comes to which the employees (new hires and old) at CDPR would rather work with.
It allows them to look for people with experience in the engine. They also won't need to do the engine work, so it will save them money. It's not a bad business decision on paper. I do think the hype behind UE5 as an engine is overblown, and the core problems of Unreal are still there.
It also won’t take them an extra 3 years to get the performance better
yes
one: it will limit them to what UE has to offer
two: it pushes UE further to a monopoly
you're not limited. developers like CDPR make their own extensions and variants to the engine. much like how Netherrealm has used UE3 right up until Mortal Kombat 1
You can change basicly everything on unreal engine with downloading their source code 😂
@@ugurinanc5177 if they plan on changing it, then why not just keep changing RED?
they can do whatever they want with the engine. They have solid engine develoepers that made red engine so they will just fix anything in UE that annoys them
Because UE is the only game engine the industry uses.
Mainly would be the cost to keep improving engine and maintain. It would cost a huge fraction of a budget.
Nanite, lumen and VSM definitely a tech CDPR looking to explore, as this tech offer function that RED Engine could not be done at current moment.
UE5 latest feature technically 'run well' with AMD which current-gen console core hardware, which is a win win situation for CDPR.
If we look deeper most of RED Engine tech are PC oriented and mainly are oriented to Nvidia GPU, for long run it is not healthy for console market.
Of course there will some flaws with UE5 but long run it better option I guess.
It's also time - I asume big part of the engine needs to be ready before game is really in developement. They can start working on new game right away and then just make changes to UE where needed.
Being able to drag items in inventory to sort it yourself as if it's like desktop icons was the coolest PC exclusive feature on Red Engine used in Witcher 3
what do you want with lumen ? UE5 can not Ray tracing. Just now RED Engine outclass UE5 in Light and shadow with raytracing.
Going from making 1 game at a time to 4 requires a lot of people… easier to hire on a widespread engine, especially to relocate in Poland.
And then, Nanite! Performances have been… a topic at launch to say the least, and Cyberpunk geometry density surely was a challenge Nanite could make easier in the future
Ever closer to a homogenized industry where everyone dev only knows how to one thing and end up in the position of Disney only having limited 3D animators. The Western game industry already collapsed ones under it's own hubris, may need to happen again for similar and different reasons. They lost devs and top people because of every other reason not related to game development at CDPR.
They've got ESG money to make up for their mistakes and keep the execs afloat while hunting for more celebrities to appear in their games. All I can think of is the generic era of PS3/XBox360 with Unreal and why that was not a good idea in the end. A industry that constantly moves towards lack of creativity, imagination, working within limitations and only knows how to do one thing can only lead to the death of it. So bizarre people think it's a good thing and blaming the engine instead of incompetence at the time rushing a project they changed direction of for mass appeal.
Speaking with such confidence about the inner workings of CDPR without a shred of evidence, makes you look like an idiot.
Your comment heavily implies that you haven't paid attention to anything CDPR has done in the last 3 years or so, (which is not surprising whatsoever) and you're reminiscing about that sweet hate train that you could ride at CP77's launch.
I think it's great they are switching to UE5. It seems that the devs want to support cyberpunk with the latest in tech for a really long time.
Maybe I'm mistaken, but from what I know about Unreal Engine the poor scalability doesn't have an easy fix. There are easier parts like using a multithreaded rendering process, but their entire Actor system is single threaded by default.
i put 150h in the game day one with no patch with my fx-8370+1060 and it fps was good
My biggest concern is frame pacing and stability. Every UE4 game I've played in the past few years have had issues with traversal stutter and eventually gets issues with crashing on multiple PCs. Every time I get excited about a game and see it's built in UE I get nervous.
UE4 here crashed as late as yesterday, because i were changing a setting in game. "Trail Out" game.
The game worked fine, but settings were buggy thanks to UE 4.
Cyberpunk crashes anyway
@@WobiKabobi no issues here with cyberpunk. Even at launch I only had 1 crash from start to finish. It crashed in the final mission after about 70 hours of play time. Haven't had a crash since replaying after the 2.0 update.
@@WobiKabobi get a better pc kid
On Linux I don’t get stutter issues
Well this discussion is not balanced without discussing UE5's advantages, most notably nanite. Nanite is an insanely advanced and cool piece of tech that they probably looked at and thought, well there's no way we can compete with that.
But at what cost? Every UE5 game/demo I’ve tried has been a colossal disappointment, due to laggy performance.
Phantom Liberty, on the other hand, looks and feels great
Doesn't matter how cool the tech is if the game stutters every five steps.
Im no software engineer or anything of the sort but from my experience playing cp2077, redengine has some horrendous draw distance and LOD issues even with mods to try and fix it
everywhere is using unreal 5 and that's good it will enable the community ultimately if that game platform takes off
I don’t understand why . Each engine has kinda unique look , and I love how red engine looks, at the end this can lead that all games will look the same if they all use unreal.
in 5 years when all these aaa ue5 games start coming out, then people will realise what kind of tech achievement this game was running today at 4k ultra 60fps + with path tracing while ue5 games wont be doing that even without basic rt effects
another thing to consider is that you might see a new console generation by the time the next game releases . the ps5 as an example is already 3 years old . it just feels more recent because of the shortages . if the next game releases in 4 years as an example that's a 7 year gap for the ps5 launch .
Pc would probably be significantly stronger too since 40 series cards by that time would be considered high tier and not epic if you get what i mean . 13th gen intel cpu's would also be thought of as old tech .
i have a 9900k . that's 5 years old and its starting to slack off in the newer titles . so much so that I'm already considering a newer gen i7 because the gains are just so much higher . but i probably wont get 13th gen because 14 is probably gonna be a thing in less than a year from now . if i had to guess . probably less than 6 months from now
Short answer : Nanite.
Nanite alone would have help a lot for the launch of Cyberpunk. They probably simply noticed that such a technology is mandatory for the kind of games they're making. With a game with the scope of Cyberpunk, LOD problems are everywhere and are still the biggest visual problems the game has now (very low textures in the background, not existing cars...) . Imagine having a technology that correct these problems by itself. Not to mention all the other features UE5 have. So, yeah, they of course thank a lot about it and took the best decision.
In fact! And with UE5.3 youll get full nanite and with that no tons of LOD settings/view distance any more. So pop ins scould be history. For big maps where you can get airborne, thats a marvel. Lumen will have path tracing too. And the newest plug ins from third parties like water or fx are just amazing. Anyway the beginning will be hard, expect the same probs like with 2077.
I think Cyberpunk 2077 pushed CDPR's RED Engine to its limit and made them realize that if they wanted to improve it for their next game, it would take more effort than moving to Unreal, on paper. (no one can tell how it would turn out yet)
There are still bugs that have persisted from day 1 of the launch. Those Bugs can be a symptom of coding flaws that RED Engine had and at this point, after many of the original RED Engine coders had already left, trying to fix those flaws within the RED Engine takes more investment than moving to Unreal.
As far as I know CDPR is working in close collaboration with epic to introduce features from the RED engine into UE5. They do detailed open worlds really well and this is what UE is still lacking. It's not made to calculate hundreds or thousand of interactable objects in their game world, one reason to why Bethesda and Larian are not using UE for their games. UE5 is not able to handle large open worlds for roleplay games, which allow you to interact with almost anything. So I hope the works CDPR does on UE5 will allow the engine to finally catch up in that regard.
" It's (UE5) not made to calculate hundreds or thousand of interactable objects in their game world"
--> EVERY well-written game engine/dev code addition will never, EVER implement a system where every possible collision/interaction with every possible entity/structure in the entire game world is calculated. For obvious reasons. Processes are 'culled' until needed ;)
@@ChrisM541 Of course they are, and UE is not good at that, because of their issues with CPU utilization. That is one of the reasons to why RPGs like Skyrim or Baldur's Gate 3 are not using Unreal engine.
@@Heruwath007 You do realise that helps my argument ;)
@@ChrisM541 Depends on what your argument is. Is it that "well-written" doesn't apply to UE? Or is it that culling does not cost anything in performance? No, wait, is it that culling is a solution for hundreds of NPCs on the screen, each of them calculating pathfinding? No worries, just cull it. Or maybe, that good devs try to optimize their engine for the game they are trying to achieve (duh), while UE is designed to drive all types of games? What is your argument?
@@Heruwath007 Lol
Unreal Engine 5 can actually run really well even on a mid range hardware provided the devs know what they are doing. Great example is The First Descendant
For this game to run on a PS4, without loading screens an on an HDD, with minimal stutters in 1.6 is out of this world.